When skeletal muscle first begins to form in the early myotome of the mouse embryo, the basic helix–loop–helix transcription factors Mrf4 (Kassar-Duchossoy et al., 2004) and Myf5 (Braun et al., 1992) play key roles as myogenic determination factors. Mrf4, but not Myf5, is also implicated in muscle cell differentiation (Bober et al., 1991; Hinterberger et al., 1991). Mrf4 is expressed in adult skeletal muscle (Hinterberger et al., 1991; Gayraud-Morel et al., 2007) and Myf5 in satellite cells (Beauchamp et al., 2000; Day et al., 2007), resident muscle stem cells. Satellite cells (Mauro, 1961) divide to provide myonuclei to the growing myofibers, before becoming quiescent in mature muscle (Schultz et al., 1978), but remain able to again proliferate and differentiate when needed for myonuclear turnover, hypertrophy, and muscle repair/regeneration (reviewed in Zammit et al., 2006).
Mrf4 is located ∼6 kb upstream of the Myf5 gene within the same locus (Fig. 1). Transcription is controlled by multiple cis-acting regulatory elements, both within the Myf5 gene, in the intergenic region and in the genomic sequence extending over ∼140 kb upstream of Myf5 (Patapoutian et al., 1993; Zweigerdt et al., 1997; Hadchouel et al., 2000, 2003; Summerbell et al., 2000; Carvajal et al., 2001; Teboul et al., 2002; Buchberger et al., 2003). This transgenic analysis has delimited regions able to direct expression at specific locations during development. In some cases, the factors that control Myf5 expression at these sites have been identified, such as sonic hedgehog/Gli and Wnt/TCF for the epaxial somite (Gustafsson et al., 2002; McDermott et al., 2005; Borello et al., 2006) or Pax3 and Six1/4 in the developing limb (Bajard et al., 2006; Giordani et al., 2007).
We have previously shown that elements located within −140 kb upstream of the Myf5 promoter also direct transgene expression to satellite cells and their progeny in adult muscle. An element able to drive robust satellite cell expression resides in the -140/88 kb interval (Fig. 1), with a second element controlling more limited expression in satellite cells located between -59 and -8.8 kb upstream of the Myf5 promoter (Zammit et al., 2004). In the case of Mrf4 transcription in adult muscle, it has been shown that ∼8 kb of sequence upstream of the rat Mrf4 gene drives myonuclear expression (Pin et al., 1997). Within this sequence, an enhancer region at -5/-4 kb (from Mrf4) that is able to direct expression in fast but not slow muscle fibers (Pin and Konieczny, 2002).
In our studies, to better define the elements responsible for transcriptional activity at this locus, we have identified the A17 enhancer, located at −17 kb from the Myf5 gene (Chang et al., 2004). This enhancer is active during development, where it shows two mutually exclusive modes of expression. Together with the Mrf4 promoter, A17 drives expression in the early embryonic skeletal muscle of the myotome in most transgenic lines. In contrast, with the Myf5 promoter, A17 transgenic lines showed later expression in fetal skeletal muscle (Chang et al., 2004). A17 directs the non-myogenic thymidine kinase (Tk) promoter to the early myotome, indicating that the enhancer itself is active in developing muscle.
We have now examined the activity of A17 transgenic mouse lines (Fig. 1) for expression in adult skeletal muscle. Neither the mouse Myf5 nor the Mrf4 promoter alone is able to drive nlacZ expression in muscle of adult transgenic mice. In conjunction with A17, however, the Myf5 promoter was targeted to satellite cells, mirroring an aspect of the expression domain of the endogenous gene (Beauchamp et al., 2000; Day et al., 2007). Expression was maintained during satellite cell activation and proliferation, but then down-regulated as the cells underwent myogenic differentiation. In contrast, with the Mrf4 promoter, A17 directed expression to myonuclei, again mimicking features of the endogenous gene (Gayraud-Morel et al., 2007). Of interest, A17 was unable to drive expression of the Mrf4 promoter in all muscles, and even all myofibers in certain muscles, highlighting the modular nature of regulation at this locus and the muscle autonomous nature of transcriptional regulation. A17 does not drive expression of the Mrf4 promoter in satellite cell-derived myoblasts until after they have begun to differentiate. These observations provide a striking example of how the two promoters can interact differently with a common upstream regulatory element to direct the distinct expression patterns of these two myogenic regulatory genes.
A17 Directs the Myf5 Promoter to Adult Skeletal Muscle Cells
Muscles were dissected from adult transgenic mice and stained overnight in X-gal solution to reveal the distribution of cells with β-galactosidase activity. It has been previously shown that transgenic lines, with the intergenic region, including the minimal Myf5 promoter, do not direct expression in adult muscle in the trunk and limbs (p8.8; Zammit et al., 2004). However, the A17 enhancer was able to drive expression of the minimal Myf5 promoter in all three A17-Myf5-nlacZ lines (52M, 54M, and 56F), with β-galactosidase activity restricted to single cells distributed throughout the muscles tested (Fig. 2a–d). Low numbers of X-gal–labeled cells were detected in the extensor digitorum longus (EDL) and soleus muscles of A17-Myf5-nlacZ mice but were very apparent in the diaphragm and intercostal muscles (Fig. 2a–d). Cryosections of diaphragm muscle showed that A17-Myf5-nlacZ transgene expression was in nuclei closely associated with myofibers, but that only a limited number of myofibers contained such nuclei (Fig. 3a). We conclude, therefore, that most satellite cells in the diaphragm express the transgene, while only a subset is positive in the EDL or soleus.
A17 Drives the Mrf4 Promoter in Both Fast and Slow Fiber Types
Two established lines and additional transient F0 transgenic Mrf4-nlacZ mice were generated, with the minimal Mrf4 promoter driving nlacZ. None of these control Mrf4-nlacZ mice expressed the transgene in adult skeletal muscle, with detailed examination of the typical line 34F, illustrated in Table 1. Six lines of A17-Mrf4-nlacZ had expression to variable degrees in adult muscle, and the two strongest expressing lines (4M and 46F) were chosen for detailed examination (Table 1; Fig. 2e–h). Expression was not uniform in all muscles within a particular A17-Mrf4-nlacZ line; however, EDL, plantaris, deep back, and temporalis muscles generally had widespread nlacZ expression in both lines, but reporter gene activity was more heterogeneous between myofibers in the soleus and serratus ventralis for example (Table 1; Fig. 2e–h).
Table 1. Comparison of Transgene Expression in Adult Muscles of Mrf4-nlacZ and A17-Mrf4-nlacZ Linesa
Expression: +++ majority of fibers, ++ many fibers, + few fibers, − none. Results are based on examination of seven adult A17-Mrf4-nlacZ line 46F mice and six of the 4M line.
Higher expression in proximal muscles than distal muscles.
Restricted to certain muscles but no proximodistal difference.
One possible explanation of the differences in these levels of A17-Mrf4-nlacZ transgene expression between muscles could be their fiber type composition, especially since in rat an enhancer has been shown to drive the Mrf4 promoter only in fast fibers (Pin and Konieczny, 2002). Mouse muscle is composed of four major fiber types, defined by their myosin heavy chain (MyHC) content as slow type I and fast type IIa, IIx, and IIb (Kelly and Rubinstein, 1994). To determine whether fiber type was a factor in the level of expression driven by A17, we examined the soleus muscle, because in mouse it contains approximately 50% slow type I and fast type IIa/IIx fibers (Agbulut et al., 2003). Immunostaining with MyHC-specific antibodies showed clearly that myonuclei with β-galactosidase activity were located in both fiber types (Fig. 3b,c). Because there is also robust transgene expression in most EDL myofibers (Figs. 2e, 3g), which mainly comprise type IIb and IIx (Rosenblatt and Parry, 1992), these observations indicate that the difference in A17-Mrf4-nlacZ transgene expression between, and certainly within, muscles probably does not reflect fiber type content per se. An A17-Tk-nlacZ line showed no adult myogenic expression that was detectable in isolated whole muscles, except occasionally in nuclei at the level of the neuromuscular junction (data not shown).
A17 Directs Satellite Cell Expression With the Myf5, but Not the Mrf4, Promoter
To better define the location of β-galactosidase activity, myofibers were obtained from the EDL, soleus, and diaphragm muscles of transgenic mice. X-gal staining revealed that individual cells situated on A17-Myf5-nlacZ-derived myofibers contained β-galactosidase (Fig. 3d–f). Their number and location indicated that they were satellite cells. In contrast, most EDL myofibers from A17-Mrf4-nlacZ mice had widespread β-galactosidase activity in myonuclei (Fig. 3g).
Immunostaining of viable myofibers isolated by collagenase digestion (Rosenblatt et al., 1995) from A17-Myf5-nlacZ mice demonstrated that β-galactosidase–positive cells also coexpressed Pax7 (Seale et al., 2001), confirming them as satellite cells (Fig. 4a–d). A17-Mrf4-nlacZ mice, however, lacked such transgene expression in satellite cells, where β-galactosidase positive nuclei and satellite cells expressing Pax7 were mutually exclusive (Fig. 4e–h).
A17 Drives the Myf5 Promoter in Satellite Cell-Derived Myoblasts
Having established the expression domain of A17-Myf5-nlacZ and A17-Mrf4-nlacZ in adult muscle, we next determined whether A17 could drive expression in satellite cells as they activate to effect muscle repair. When myofibers are cultured, the associated satellite cells will migrate from myofibers onto the tissue culture substrate where they proliferate, differentiate, and fuse into myotubes (Rosenblatt et al., 1995). Myofibers were isolated from EDL and soleus muscles of A17 transgenic mice, plated, cultured, and assayed for nlacZ expression by X-gal coloration. Two time points were examined: 72 hr to investigate transgene expression in proliferating satellite cell-derived myoblasts and 7 days to analyze expression after differentiation.
After 72 hr in culture, A17-Myf5-nlacZ-derived EDL and soleus myofibers were surrounded by many satellite cell-derived myoblasts with nuclear β-galactosidase activity (Fig. 5a–d). By 7 days in culture, many myoblasts still exhibited high β-galactosidase activity, whereas in general, the nuclei of differentiated myotubes showed less activity or were negative (Fig. 5a–d). This finding probably reflects perduration of β-galactosidase in myotubes in which transgene expression was down-regulated on differentiation, as seen for Myf5 transcription in muscle cell culture (Montarras et al., 1989). All EDL (n = 83 and n = 156, two mice) and soleus (n = 40, one mouse) myofibers examined had associated satellite cells able to express the A17-Myf5-nlacZ transgene when activated. In contrast, the transgene composed of the intergenic region containing the Myf5 promoter does not drive nlacZ expression in either satellite cell-derived myoblasts or myotubes (p8.8; Zammit et al., 2004).
A17 Directs the Mrf4 Promoter to Satellite Cell-Derived Myotubes
A17-Mrf4-nlacZ myofibers retained reporter gene expression in myonuclei of isolated adult EDL and soleus myofibers in culture. Examination of cultures after 72 hr with X-gal coloration showed that proliferating satellite cell-derived myoblasts did not have β-galactosidase activity (Fig. 5e–h). After 7 days in culture, however, extensive β-galactosidase activity was observed in the nuclei of satellite cell-derived myotubes of A17-Mrf4-nlacZ line 4M, indicating that the transgene was activated during differentiation (Fig. 5e–h), consistent with its expression in myonuclei of adult muscle.
In contrast, Mrf4-nlacZ lines failed to show any β-galactosidase activity in either myofibers, satellite cell-derived myoblasts, or myotubes after culture for 72 hr and 7 days (Fig. 5i,j). To determine the specificity of the A17 enhancer, isolation, culture, and X-gal coloration of myofibers from A17-Tk-nlacZ was performed. This analysis revealed that occasional myofibers gave rise to very limited numbers of myoblasts with nlacZ expression after 72 hr and 7 days of culture (Fig. 5k,l).
The 583-bp A17 enhancer sequence drives transgenic expression from the Myf5 and Mrf4 promoters in two distinct patterns in skeletal muscle during development. Expression is either found in the developing myotome during the embryonic period, but is then down-regulated during the fetal period, or expression is initiated in the fetal period (Chang et al., 2004). This first mode, usually seen with the Mrf4 promoter, reflects the early myotomal expression of this gene. In general, fetal expression is directed by the Myf5 promoter, suggesting that A17 is also active during this later wave of Pax3/7-dependent myogenesis (Relaix et al., 2005). Interestingly, the A17 transgenic lines that have the fetal pattern of expression also generally exhibit expression in adult muscle, demonstrating a link between control of fetal and adult myogenesis.
Because A17 was isolated in an enhancer trap assay from the C2 myogenic cell line, originally derived from postinjury adult (2 month) thigh muscle (Yaffe and Saxel, 1977), it is not surprising that it is also active in adult muscle. A17 linked to the Myf5 or Mrf4 promoters can reproduce aspects of the endogenous expression of the two myogenic regulatory genes in adult muscle. Strikingly, the site of expression depends on the promoter. With the Myf5 promoter, A17 drives activity in a subset of quiescent satellite cells and myoblast progeny. In contrast, A17 directs the Mrf4 promoter in myonuclei of differentiated muscle fibers and in satellite cell-derived myotubes, but not myoblasts. These observations are consistent with those on transgenes with the rat Mrf4 promoter (Pin et al., 1997). The role of A17 in driving these different expression profiles is highlighted by the observation that mice transgenic for either the Myf5 or Mrf4 promoter show no activity in adult muscle. Furthermore, muscle expression can be detected with the A17 element and a non-myogenic Tk promoter during development (Chang et al., 2004). Although there was no significant expression in adult skeletal muscle of A17-Tk-nlacZ transgenic mice, when cultured, rare individual satellite cell-derived myoblasts with transgenic activity could be detected both at 72 hr and 7 days of culture, conferring a pattern of expression on the Tk promoter similar to that of A17-Myf5-nlacZ. This finding indicates that A17 itself may have a degree of satellite cell specificity. The question of how these two promoters interact differentially with A17 and other regulatory elements of the locus is of major importance in understanding the different expression profiles and myogenic roles of Myf5 and Mrf4.
Functional analysis of A17 demonstrated that a core 15-bp sequence containing an E-Box is essential for activity, since its deletion resulted in a loss of developmental muscle expression (Chang et al., 2004). This E-box binds USF, and overexpression of a dominant negative form of USF1 or USF2 represses A17 activity in C2 cells (Chang et al., 2004), which were originally derived from adult muscle (Yaffe and Saxel, 1977). USF1 and USF2 are ubiquitously expressed in the adult, but neither USF1 nor USF2 null mice display a muscle phenotype. Double homozygotes die in utero, however, suggesting redundant functions (Sirito et al., 1998).
Identification of the sequences that regulate Myf5 expression at different sites of myogenesis in the embryo revealed the remarkable extent to which spatiotemporal control is exerted (Buchberger et al., 2003; Hadchouel et al., 2003). In the adult, we have previously shown that elements in the -140 kb to -88 kb interval (Fig. 1) drive widespread expression from the Myf5 promoter in satellite cells (Zammit et al., 2004). A second element that is also able to direct transgene expression to satellite cells, is located between -59 and -8.8 kb (Zammit et al., 2004), a region that encompasses the A17 element (Fig. 1). Deletion of A17 in the context of a YAC containing -240 kb of upstream sequence with both Myf5 (nlacZ) and Mrf4 (alkaline phosphatase) targeted by reporter transgenes did not result in an obvious diminution in expression, suggesting compensation by other upstream regions. More extensive deletion experiments with all regions implicated in expression in adult muscle will be necessary to further clarify their relative roles.
In the case of Mrf4 expression in the myonuclei of muscle fibers, it is also probable that several elements are involved. With the Mrf4 promoter, only some muscles are targeted by A17. The A17 sequence, which shows 76% conservation between rat and mouse, lies at the 5′ extremity of the 8.5 kb of rat genomic DNA, which has been described as giving extensive expression in adult skeletal muscle (Pin et al., 1997). The A17 enhancer in conjunction with the Mrf4 promoter does not preferentially target the myonuclei of slow or fast fibers, unlike the enhancer element at -5/-4 kb from the rat Mrf4 gene, that is only active in fast fibers (Pin and Konieczny, 2002). These reports on the rat, together with what we observe for A17 with the Mrf4 promoter in the mouse, again point to a complex modular regulation for the myonuclear expression of this gene.
The Mrf4/Myf5 locus provides an interesting paradigm with which to study regulatory mechanisms. Despite their close vicinity, Myf5 and Mrf4 are differentially regulated in time and space. The activity of regulatory regions has been described in the context of both genes (Carvajal et al., 2001), but more detailed analysis of enhancer elements has largely been limited to Myf5. Identifying and characterizing the adult regulatory elements of the Mrf4/Myf5 locus will give new insight into the underlying regulatory strategies that govern skeletal muscle homeostasis and regeneration at different sites in the body. The range of human myopathies, in which skeletal muscles are differentially affected by mutations, already points to the extent of skeletal muscle heterogeneity. As in the case of the embryo, this is not only situated at the level of fiber type and physiological function, but also concerns the regulatory factors and how they in turn are regulated. The A17 enhancer provides another example of a transcriptional control element in the Mrf4/Myf5 locus that can direct specific temporal and spatial expression of each promoter.
Transgenic mice were generated by microinjection of purified DNA fragments into fertilized (C57BL/6J X SJL) F2 eggs as described previously (Hadchouel et al., 2000). Injected eggs were then reimplanted the day after injection into pseudopregnant (C57BL/6J X CBA) F1 foster mothers. Transgenic mice were identified by polymerase chain reaction of DNA prepared from mouse tails using a primer situated in lacZ paired with a primer situated in the promoter of either Myf5, Mrf4, or Tk.
Mrf4-nlacZ lines consist of the minimal Mrf4 promoter (-385 bp from the ATG; Black et al., 1995) driving nlacZ, of which two lines were established, with several transient transgenics additionally examined. The 5′ addition of the 583-bp sequence of A17 was used to generate A17-Mrf4-nlacZ (Fig. 1). A total of 12 lines of A17-Mrf4-nlacZ were established, of which 3 lines failed to demonstrate transgene expression, 3 lines showed only developmental expression, and the remaining 6 had variable expression in adult muscle, the 2 strongest expressing of which (4M and 46F) were chosen for detailed examination. The A17-Myf5-nlacZ lines consisted of the 583-bp sequence of A17, the proximal branchial arch enhancer of Myf5 situated -1.8 to -0.7 kb upstream of the Myf5 ATG (as an internal positive control for embryonic expression) and the minimal Myf5 promoter (-291 bp from the ATG; Summerbell et al., 2000) controlling nlacZ (Fig. 1). Three A17-Myf5-nlacZ lines were established (Chang et al., 2004) with expression levels in line 56F higher than in 52M and 54M, and so it was mainly used for analysis, although there were no major qualitative differences in nlacZ expression profiles between lines. All transgenic lines analyzed had a low transgene copy number. Experiments on mice were carried out in accordance with French law.
Tissue Preparation, Myofiber Isolation, and Culture
Adult mice were killed by cervical dislocation and several muscles were removed, including the EDL, tibialis anterior (TA), soleus, plantaris, diaphragm, masseter, and intercostal muscles. These were then either fixed in 4% paraformaldehyde/phosphate buffered saline (PBS) for X-gal coloration, snap frozen for cryosectioning, or digested to isolate single myofibers.
The isolation of viable myofibers has been described in detail elsewhere (Rosenblatt et al., 1995). Briefly, EDL and soleus muscles were incubated for 80 min at 37°C in 0.2% (w/v) type 1 collagenase in Dulbecco's modified Eagle's medium (DMEM; Gibco) with 2 mM L-glutamine (Sigma) and 1% penicillin/streptomycin solution (Sigma). Myofibers were then isolated by trituration and either fixed in paraformaldehyde/PBS or cultured.
To determine whether satellite cell-derived myoblasts had transgene activity, myofibers were cultured under conditions where they adhere to the substrate and satellite cells migrate away from the fiber and proliferate (Rosenblatt et al., 1995). Myofibers were placed in 24-well plates (Marathon) coated with 1 mg/ml Matrigel (Collaborative Research Inc.). Plating medium (DMEM with 10% [v/v] horse serum; PAA Laboratories) and 0.5% (v/v) chick embryo extract (ICN Flow) were added, and the myofibers were incubated at 37°C in 5% CO2. Myofibers and cells were fixed in 2% paraformaldehyde/PBS for 2–5 min from 72 hr to 7 days after plating.
To assess the distribution of cells with β-galactosidase activity, fixed muscles, myofibers, and cells were incubated in X-gal solution (4 mM potassium ferrocyanide, 4 mM potassium ferricyanide, 2 mM MgCl2, 400 μg/ml X-gal, and 0.02% NP40 in PBS) at 37°C overnight. Myofibers were then rinsed several times in PBS and mounted in DakoCytomation Faramount aqueous mounting medium. Images were captured by a Zeiss Axiocam CCD camera using Zeiss Axiovision 3.0.6 SP3 software.
Fixed myofibers were permeabilized with 0.5% (v/v) Triton X-100/PBS and blocked by incubation in 20% (v/v) goat serum in PBS for at least 30 min. Primary antibodies used were monoclonal mouse anti–β-galactosidase (clone 40-1A, Developmental Studies Hybridoma Bank), anti-Pax7 (Developmental Studies Hybridoma Bank), anti-slow MyHC (clone NOQ7.5.4D, Sigma), anti-desmin (clone D33, DakoCytomation), and rabbit polyclonal anti–β-galactosidase (Molecular Probes). Primary antibodies were applied overnight at 4°C and then visualized by incubation with Alexa Fluor-conjugated secondary antibodies (Molecular Probes). During the final PBS wash, slides or dishes were stained in 1 μg/ml Hoechst 33342 before mounting in PBS/75% glycerol, or else mounted in DakoCytomation Faramount fluorescent mounting medium containing 100 ng/ml 4,6-diamidino-2-phenylindole (DAPI). Immunostained myofibers were viewed on a Zeiss Axiophot epifluorescence microscope. Digital images were acquired with a Charge-Coupled Device (RTE/CCD-1300-Y; Princeton Instruments Inc.) at −10°C using Metamorph software, version 4.5r5 (Universal Imaging Corp).
We thank James Minchin and Simon Hughes for assistance and access to microscopes, the colleagues who made their antibodies available through the Developmental Studies Hybridoma Bank and the CIGM of the Pasteur Institute for helping to generate transgenic lines. T.C. was funded by an NIH NRSA postdoctoral fellowship, a fellowship from the Association Française contre les Myopathies (AFM), and a Human Frontiers Science Program grant. S.V. is an INSERM research fellow. The laboratory of P.Z. is supported by The Medical Research Council UK, The Muscular Dystrophy Campaign, and the Association of International Cancer Research. The laboratory of M.B. is supported by the Pasteur Institute and the CNRS, with additional grants from the AFM. We acknowledge the support of the MYORES Network of Excellence, from the European Commission 6th Framework Programme to P.Z. and M.B.